1. Field
Apparatuses and methods consistent with exemplary embodiments relate to a watermark processing device and method, and an image capturing apparatus capable of reducing a processing time and a calculation amount while inserting a watermark into an image
2. Description of the Related Art
Due to rapid development of the Internet, transferring and storing digital data has become easier and faster via a wired/wireless network. However, the digital data is so easily copied and modified that it is difficult to discriminate original data from a duplicate. As illegal duplication and distribution spreads, intellectual property rights have arisen over time as a crucial social issue. In order to address this issue, digital watermarking is being researched and practically applied as a technology to serve a basis to protect intellectual property rights and claim a copyright. In early 1990s, applying watermarking to digital images began to be researched. As the watermarking mainly researched in a spatial domain in the early days is now being applied to a frequency domain, its research domain has been changing. Although frequency domain watermarking is robust to an attack compared to spatial domain watermarking, it is hard to precisely select a watermark insertion position due to frequency characteristics. However, the watermark insertion becomes more efficient once a wavelet domain having both frequency characteristics and spatial characteristics is introduced.
A watermark insertion method using a wavelet transform includes performing a 3-level wavelet transform on an image, inserting a watermark, and then performing an inverse wavelet transform. The inserted watermark may be extracted by performing a wavelet transform on an image having a watermark therein and calculating correlation coefficients from inserted level image information and a reference watermark. A watermark insertion/extraction method is robust even in a poor communication environment and during codec compression. It is difficult to remove the watermark.
In order to implement a typical watermark insertion method by using a wavelet transform, an image is required to be wavelet-transformed and finally inversely wavelet-transformed again. In this case, since a two-level wavelet transform is normally adopted, a calculation amount becomes larger. Even when watermark information to be inserted is a fixed value, it takes a long processing time to perform a wavelet transform and then insert a watermark in a wavelet domain. Especially, since the number of pixels of images obtained by a surveillance camera has recently increased to be in the megapixel range, it is difficult to apply watermark insertion to a low specification surveillance camera due to limitations in a calculation amount and a processing time.